LEDs in parallel -no problems? 1

[grigson]

Hello,

My company has been putting 1A power LEDs from Cree and Osram in parallel for soome time now.

We only ever use a single current source, and NEVER use series current equalisation resistors or current mirrors.

We have encountered no problems......so why is everyone telling us we are doing wrong?...i just dont understand
A batch of the same type of power LEDs always have virtually exactly same Vf's, and so no LED hogs the current.
Driving paralleled LEDs with no equalisation resistors is the cheapest way to drive LEDs.....you can just usually use a cheap simple hysterwetic buck converter.
Why are so many people (even some prospective customers) telling us we are doing wrong....we have had no problems.
On our production line, each finished product gets left ON for 20 minutes before being bagged and despatched..........if a failure was going to occur, it would occur there...but it never has, so we are thinking of doing away with this mini soak test.
Are LED lights are in the power range of 2W to 50W.

Why is what we are doing wrong?
Do you agree with us that this talk of unequal Vf's is a myth?

 

[IRstuff]

Vf is highly dependent on the LED manufacturing process, and just because Osram and Cree have their sh!t together and run a tight process does not mean that your position is correct. You are just lucky. Since they can and most likely will change their process to suit their own needs, it's possible that that you'll be blowing up LEDs in the future.

The bottom line is that you're ignoring the manufacturer's datasheet at your own peril.

http://www.cree.com/~/media/Files/Cree/Chips and Material/Data Sheets Chips/CPR3ES.pdf

clearly stipulates a Vf range of about 0.8V. All else being equal, it only takes 59mV to get a 10x change in forward current, so that's what you're betting on. Now, you are probably lucky in that the current regime that you're operating in is partly dominated by parasitic resistances, so the variance in forward voltage is swamped out by the parasitics.

TTFN
faq731-376

 

[btrueblood]

What IRstuff said. I would add two more points: 1. that the Vf decreases with increasing temperature, i.e. if one LED does start to hog current, it will need to dissipate more heat, which raises the junction temperature, which decreases Vf causing more current to flow...resulting in a thermal runaway. 2. What happens if one LED goes poof - does your current source compensate? If not, are you suddenly dumping n/(n-1) more current through the remaining resistors, and how do you think that will affect #1?

 

[MacGyverS2000]

On our production line, each finished product gets left ON for 20 minutes before being bagged and despatched..........if a failure was going to occur, it would occur there...but it never has, so we are thinking of doing away with this mini soak test.

That's some of the worst testing metrics I've seen in a long time. 20 minutes is a flyspec. Under what conditions are these tests performed? If it's a well-A/C'ed room with good airflow, that test could very well pass every time... put it under a cabinet with low airflow and those LEDs may not last more than 100 hours.

You need to stop ignoring the combined wisdom of the companies who make (and test!) this stuff simply because you can't make it fail with a 20-minute test. The moment you get an LED in the batch with a Vf more than a tenth of a volt or two out of nominal, you'll see the entire stack fail in short order. If I was a customer and saw that design, I'd look elsewhere for lighting components.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Comcokid]

If the parallel arranagement of LEDs is working, then I don't see much of a problem.

Your circuit from your other post only has two LEDs in series on each string, plus a diode with six diode/LEDs strings in parallel. If one string draws slightly more current, then the drop across the diode will increase. The diode is functioning somewhat like the resistor that might usually be used.

Also, many of the newer LED driver buck-type ICs will show parallel strings of series LEDs. Also, Osram does extensive grading of their LEDs during the manufacturing test which will keep all the LEDs in a ordered lot the same. (But, you never can get the identical part number from lot to lot as the grading suffix of the part number changes.) You should not have an issue as long as you don't mix LEDs from differnt lots.

 

[ScottyUK]

ComcoKid,

Same batch? They're not even the from the same manufacturer! See the initial post.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[OperaHouse]

I have a 10W LED which internally is 4 strings of three in sries. Obviously the LED manufacturer doesn't have a problem with this. With multiple LED in series, the internal resistances add up and will load share better in parallel operation. Large single parallel operation is dicier as you operate near max current.

I will say reverse voltage spec of typically 5V seems to be a worthless holdback. I've never found a LED that wasn't good for over 200V reverse voltage. I've done a test for several months with no degredation and with units that were 30years old.

 

[mcgyvr]

Redesign for series strings, use ballast resistors in the parallel strings or keep spinning that chamber in the game of Russian Roulette.
Hire a real engineer first.. Even a 1st year student could tell you how wrong it is.

 

[VEBill]

...even some prospective customers telling us we are doing wrong...

How would prospective customers even find out about such a thing? Strange.

If it's any consolation, the typical inexpensive Xmas Tree LED strings defy formal analysis too.

 

[MikeHalloran]

Operating outside the parameters given in the spec sheet is not "wrong".

It's a business decision.

From a technical perspective, which is mostly what you get here, it's fluorescently stupid.

... but I'm sure some MBA will be able to defend the decision.

... or to blame some poor technical schmuck when the peanut butter hits the fan.



Mike Halloran
Pembroke Pines, FL, USA

 

[LionelHutz]

Do you agree with us that this talk of unequal Vf's is a myth?

No!

 

[signal2noise]

I don't think the unequal Vf's is a myth. Much of it has to do with the thermal coupling between the two LED's and the resistance between each one and its common point. Low resistance and poor thermal coupling makes this unstable and much more likely to result in thermal runaway.

Why would the Vf theory be a myth when it's something so simple do disprove.